Le 17 mai 2016, 10:47 dans Humeurs • 0
Now social network become more and more important in people’s life, it help us collect a lot of friends and come closer with them. The rich network bring us great information benefits, we can get some knowledge that we never know, and we can obtain some import message we need, such as job opportunity, something on sales, and traveling experience.
There have some popular communication tools like facebook, linkedin and twitter. Many people create their social network through these platforms, for the tools help them instantly communicate with their friends, and share something with others. However, the business cards are till popular in people’s life, especially on business behave, people always use paper business cards to exchange their personnel information. Some cards contain 2-D barcode, and we can easily scan the code to fast enter the information and add the card owner to our friend list. But most of the cards do not have 2-D barcode, how to quickly save the information?
The Yunmai Business Card Recognition SDK can do that. This (Optical Character Recognition) OCR engine is special for reading business card, it is able to extract a business card information form the card image within 1 second, and the recognition accuracy rate reach 99%. It can be integrated with PC software and mobile applications. As long as the software obtain the image of the card, the Yunmai Business Card Recognition engine will read the card information. Therefore it is good choice for developers to use this OCR SDK to add business card recognition feature into their software.
The process of OCR engine
- Get image
- Blur detection
- Text lines segmentation
- Chracters segmentation
- Chracter feature extraction
- Chracters matching
- Handling End of line
- Automatic words segmentation
- Chracters detection
Accuracy rate and OCR engine efficiency (reference)
- Character recognition accuracy
The following are the average recognition accuracy for the business cards of different languages:
English characters: 97%
European characters: 89%
Chinese characters: 95%
These data are obtained on recognizing different format and font size business cards.
- BCR engine efficiency
The engine process the business card recognition will take about 0.6 seconds for PC per card for average.
Programming language supported
The BCR SDK is available for different programming languages: Java, C++, C, Object Pascal, Objective-C.
Other recognition SDKs we developed
Document OCR SDK
Bank card recognition SDK
Chinese citizen ID card recognition SDK